AI/ML Google Chrome silently downloads large AI model, raising privacy concerns May 8, 2026 Share By SC Staff As detailed in HackRead, a cybersecurity researcher has uncovered that Google Chrome is downloading a substantial AI model, Gemini Nano, onto users' computers without explicit consent. This discovery raises significant questions about user privacy and the energy consumed by such background data transfers. Cybersecurity researcher Alexander Hanff discovered that Google Chrome is downloading a 4GB AI model, Gemini Nano, onto users' machines if they meet certain hardware requirements. The download occurs silently in the background during idle times, initiated by the browser itself without any user interaction. Hanff's tests, conducted on a clean Chrome profile on macOS in April 2026, confirmed the browser created an "OptGuideOnDeviceModel" folder and downloaded the "weights.bin" file within 14 minutes. Google states the model aids in features like scam detection, but the method of distribution raised concerns among critics. Hanff estimates that if 100 million users download it, the energy consumption would be 24 GWh, escalating to 240 GWh if it reaches 30% of Chrome's user base. Furthermore, Hanff argues this silent installation potentially violates the EU ePrivacy Directive and GDPR, which mandate transparency and consent for data storage on user devices. He cited breaches of Article 5(3) of the ePrivacy Directive and GDPR principles, including lawfulness, fairness, transparency, and data protection by design. A similar issue was noted with Anthropic's Claude Desktop app. Users can check for the model at "chrome://on-device-internals" and disable it via Chrome Settings or "chrome://flags". Source: HackRead An In-Depth Guide to AI Get essential knowledge and practical strategies to use AI to better your security program. Learn More SC Staff Related Security Operations Tanium and ServiceNow partner for autonomous IT operations SC Staff May 8, 2026 The collaboration merges Tanium's real-time endpoint intelligence with ServiceNow's workflow orchestration to address the gap between IT visibility and action. Phishing Herd Security raises $3 million for AI-powered security training SC Staff May 8, 2026 Founded in 2025, Herd Security provides an agentic AI platform designed for continuous security awareness training. Security Operations Boost Security acquires 2 startups, raises $4 million for AI defense platform SC Staff May 8, 2026 Boost Security has acquired SecureIQx, an MIT-founded startup specializing in software composition analysis reachability, and Korbit Technologies, an AI-driven pull request review platform. Get daily email updates SC Media's daily must-read of the most current and pressing daily news Business Email By clicking the Subscribe button below, you agree to SC Media Terms of Use and Privacy Policy . Subscribe You can skip this ad in 5 seconds